The Prediction Model of Bankruptcy: Evidence from the Small and Medium Enterprises (SMEs) in Thailand

Transcription

1 Vol. 3, No. 10, 2014, The Prediction Model of Bankruptcy: Evidence from the Small and Medium Enterprises (SMEs) in Thailand Yossavadee Pugpaichit 1, Phassawan Suntrauk 2 Abstract The study aims to develop the bankruptcy prediction model of Small and Medium Enterprises (SMEs) in form of company limited in Thailand during Using logistic regression analysis and in-the-sample data,results show that the bankruptcy prediction model consists of a ratio of earnings after taxes to total assets and an asset turnover ratio. Since these two financial ratios represent profitability and asset utilization of firms, it is asserted that bankrupt firms are those who have relative low profitability due to their inefficient use of assets in generating profits continuously. By using out-of-sample data to examine the predictive ability of the estimated model, the results reveal that the estimated prediction model provides favorable results in which the percentages of accuracy of predicting bankrupt and non- bankrupt firms are 68% and 60%, respectively. Keywords: Bankruptcy, Financial Ratios, SMEs, Thailand, Logistic regression Classification: JEL: G33 1. Introduction The well-known financial crises such as the great depression in 1929, the Asian financial crisis in 1997, and the subprime mortgage crisis in 2007 are examples of severe financial problems, causing thousand firms around the world to become bankrupt. These crisis events have raised the awareness of financial analysts and academic researchers to realize the importance of bankruptcy prediction. Since the late 1960s, there have been many studies regarding bankruptcy prediction. Beaver (1966) attempted to identify whether the financial ratios are able to predict the probability of bankruptcy. He revealed that the financial ratios can signal useful information regarding financial conditions of firms well before the bankruptcy. Moreover, the financial ratios of bankrupt firms significantly differ from those of nonbankrupt firms. Besides, Altman (1968) developed a multivariate discriminate model which consists of five ratios representing profitability, liquidity, solvency, leverage, and activity of firms. He pointed out that the significant variable that can classify the bankrupt firms from those who are not is the profitability ratio, followed by activity ratio, and solvency ratio. In addition, Deakin (1972) indicated that the financial ratios can be used to predict firms bankruptcy at least three years in advance. Furthermore, Ohlson (1980) applied the logit analysis to predict the probability of bankruptcy. He examined the ability of eight financial ratios and firms sizes in predicting the probability of bankruptcy. He found that financial ratios, which are total liabilities to total assets ratio, net income to total assets ratio, and working capital to total assets ratio properly predict the probability of bankruptcy one year in advance. The studies aforementioned have prompted many researchers to further investigate the roles of financial ratios in the prediction of bankruptcy. For instance, Kantong (1997) found that in case of Thailand, the 1 Master of Science in Finance and Economics Martin De Tours School of Management and Economics, Assumption University Hua Mak Campus, Bangkok Thailand 2 Business Administration Division, Mahidol University International College Salaya Campus, Nakorn Pathom Thailand 2014 Research Academy of Social Sciences 788

2 financial ratios, such as the retained earnings to total assets ratio, operating income to total assets ratio, total liabilities to total assets ratio, are the potential predictors of firms failure. In addition, Nam and Jinn (2000) included thirty-three financial ratios when developing the model as an early warning sign of bankruptcy of firms. Their investigations included the financial ratios that measure profitability, turnover, growth, productivity, fixed charge coverage, solvency, leverage, and liquidity. Furthermore, Charitou et al. (2004) who used a large number of financial ratios, proved to be successful in predicting the occurrence of firms bankruptcy in prior studies, in their bankruptcy research.recently, Suntraruk (2010) revealed that besides nonfinancial factors, financial ratios which are the retained earnings to total assets ratio and the asset turnover ratio can be used to predict the probability of bankruptcy in case of nonfinancial firms listed on the Stock Exchange of Thailand. As the matter of fact, financial ratio analysis is beneficial in several perspectives. Financial ratios work as the mirror to reflect financial situation and performance of firms. They are used by stakeholders to assess the strengths and weaknesses of firms, to be as a guideline in order to improve the operation s weaknesses within firms, to identify the ability of firms to settle debts and obligations,to evaluate stability of firms in order to reduce chance of losing from investment, and to predict the future financial events in order to decrease chance of future financial bankruptcy (Ingram, 2009). Since the financial ratios are important, the primary objective of this study is to develop the bankruptcy prediction model by using the financial ratios as the predictors. The study, however, emphasizes on the firms that are classified as Small and Medium Enterprises (SMEs) in the form of company limited in Thailand. These firms are the one that are too big to be just small firms, but are too small to be regarded as big firms. In case of Thailand, to classify a firm as small or medium enterprise, there are three criteria used which are types of business, number of employees, and value of fixed assets. Table 1 presents the selection criteria of Small and Medium Enterprises(SMEs). Table 1: Classification Criteria of SMEs in Thailand Sectors Number of Employees Value of Fixed assets (million baht) Small Medium Small Medium Manufacturing < < 50 > Services < < 50 > Trade: Wholesale < < 50 > Retail < < 30 >30-60 Source: The Office of Small and Medium Enterprises Promotion (OSMEP, 2011a) According to the white paper on Small and Medium Enterprises of Thailand published by the Office of Small and Medium Enterprises Promotion, By year 2010, total Small and Medium Enterprises(SMEs) in Thailand was million firms, which was accounted for 99.6% of businesses in Thailand. Moreover, the number of employees employed in this segment was 77.86% of overall employment in country (OSMEP, 2011b). In addition, according to the database of Business Online Public Company Limited (BOL), the following table 2 shows that from year 2005 to 2010, the number of Small and Medium Enterprises (SMEs) became increasing. All of these indicate that Small and Medium Enterprises (SMEs) is one of important businesses that drive Thai economy. In Thailand, there exist three sectors in the Small and Medium Enterprises (SMEs). In year 2010, the most important sector was manufacturing, following by service and trade sectors since it was reported that the market values of all final goods and services produced by these three sectors were 32.3%, 31.6% and 28.2% of all SMEs market values, respectively(osmep, 2011b). 789

3 Year Y. Pugpaichit & P. Suntrauk Table 2: Number of all SMEs and SMEs in Manufacturing Industry during Number of SMEs Number of SMEs in Manufacturing Industry Number of Firms Percentage (%) ,249, , ,287, , ,375, , ,836, , ,900, , ,924, , Source: Annual Report of the Office of SMEs Promotion during Although manufacturing industry is key sector in this type of business,comparing to other SMEs sectors, more than 1,000 limited companies in manufacturing sector went bankrupt in 2005 to Furthermore, as shown in table 2, the number of SMEs in manufacturing sector has decreased since 2005, implying that the number of bankrupt firms continued rising. As the failure of firms in this manufacturing sector is a great loss to not only SMEs industry, but also Thai economy, the primary objective of study is to develop a bankruptcy prediction model of SMEs manufacturing firms in Thailand.This warning tool would help stakeholders make correct decision, prevent firms from bankruptcy, and finally help reduce losses from this severe failure event. The paper is organized as follows. The next section discloses prior studies regarding the relationship between financial ratios and bankruptcy. The third section explains the data collection and logistic regression model. The fourth section reports empirical results. The last section provides conclusion. 2. Prior Studies Financial ratios can be used to measure the firm performance in four perspectives; liquidity,leverage, asset utilization, and profitability.the following explains the relationship between financial ratios and bankruptcy mentioned in prior studies. Liquidity Liquidity ratio indicates how effectively a firm is able to meet its short-term obligations, e.g. daily expenses, the payment of bills, periodic interest due, etc.a firm having high liquidity is able to generate sufficient cash flow to pay off its short-term obligations(brigham & Ehrhardt, 2005).For this study, in accordance with Altman (1968), a ratio of net working capital to total assets is used to measure the ability of firm to match its liquid net assets with its short-term liabilities. A high ratio is favorable and in general, healthy firms tend to have a high ratio. With a high ratio, the firm is able to remain solvent during downturns. On the other hand, a firm having financial problems is more likely to have a relatively low ratio because of the risk of insolvency in the short-term, causing another doubt regarding its long-term performance.hence, the negative relationship between a ratio of net working capital to total assets and bankruptcy is expected. Many researchers support this negative relationship. Merwin (1942) examined the predictors of bankruptcy by using a sample 200 firms from five industries that are banking, clothing for men, furniture, stone-clay, and machine tool industries. He concluded that a ratio of net working capital to total assets is one out of three variables used for bankruptcy prediction. In addition, using the Multiple Discriminant Analysis (MDA) with a sample of 66 companies, Altman (1968) concluded that a ratio of net working capital to total assets has a negative relationship with bankruptcy. In addition, Grammatikos and Gloubos (1984)revealed 790

4 that for the bankrupt firms, a ratio of net working capital to total assets is likely to decrease at least one year before the crisis happened. Leverage Leverage ratio signals how a firm finances its investment with debts. Due to the tax burden benefits and the absence of share dilution, many firms prefer debt financing to capital financing (Brigham & Ehrhardt, 2005).Nonetheless, Kraus and Litzenberger (1973) asserted that a high degree of debts increases bankruptcy costs. A firm having high degree of debts might not survive for many years ifit fails to pay off its obligations as a consequence of a decrease in future cash flow. Hence, the leverage ratio can be used as a predictor of firms bankruptcy. For this study, the firm s leverage is proxied by a ratio of total liabilities to total assets. According to Beaver (1966),a ratio of total liabilities to total assets is one significant variable that can be used to predict bankruptcy. Indeed, he found a positive relationship between a ratio of total liabilities to total assets and the chance of bankruptcy. Besides,Ohlson (1980) applied logistic regression analysis and revealed that the relationship between total liabilities to total assets ratio and the chance of bankruptcy is positive. In addition, Fulmer et al. (1984) used step-wise multiple discriminant analysis to evaluate forty financial ratios and he reported that a ratio of total liabilities to total assets is positively related to chance of bankruptcy.further more, in case of Thailand, Kantong (1997) asserted that a ratio of total liabilities to total assets relates to bankruptcy. In summary, an increase in a ratio of total liabilities to total assets reduces the performance of firm. This can further increase the chance of bankruptcy. Hence, a positive relationship between a ratio of total liabilities to total assets and chance of bankruptcy is hypothesized in this study. Asset Utilization Asset utilization ratio can explain the ability of a firm in utilizing its assets to generate revenue efficiently. In this study, consistent with Altman (1968),an asset turnover ratio is a proxy of asset utilization. A firm that is able to manage its assets effectively tends to produce high revenue. This healthy firm is more likely to have high asset turnover ratio and less likely to face insolvency. On the other hand, firm having low asset turnover ratio is those having problems in utilizing its assets, such as cash, accounts receivable, inventory, fixed assets, etc. to generate revenue. The revenue of such firm is relative low. Hence, a firm having low asset turnover ratio is more likely to face difficulty in paying off its debts, particularly during the business downturn. Several researchers use this asset turnover ratio to predict bankruptcy. Altman (1968) and Fulmer et al. (1984) revealed that total asset turnover ratio is a significant independent variable for bankruptcy prediction. They found out that there exists a negative relationship between total asset turnover ratio and the chance of bankruptcy. Moreover, Springate (1978) employed a stepwise multiple discriminant analysis and asserted that total asset turnover ratio is one out of four financial ratios which is beneficial for prediction of bankruptcy. He reported a negative relationship between total asset turnover ratio and the chance of bankruptcy. Furthermore, Suntraruk (2010) revealed that the probability of bankruptcy of listed firms in Thailand reduces with an increase in total asset turnover ratio. In conclusion, this study hypothesizes that there exists a negative relationship between total asset turnover ratio and the probability of bankruptcy.an increase in total asset turnover leads to higher performance of firm and reduces chance of bankruptcy. Profitability Profitability ratio is basically one of significant ratios indicating the performance of a firm (Altman, 1968). Itprovides an assessment of firm s effectiveness in generating profits. Altman (1968) found out that high retained earnings to total assets ratio, a proxyof profitability, is desirable because it shows the ability of firms to increasingly retain more profits.in addition, Beaver (1966)supported the negative relationship between a ratio of earnings after taxes to total assets and bankruptcy. A decrease in a ratio of earnings after 791

5 Y. Pugpaichit & P. Suntrauk taxes to total assets increases the chance of bankruptcy. Moreover, Suntraruk (2010) reported that the retained earnings to total assets ratio, a measure of profitability, is negatively related to the probability of bankruptcy. Listed firms in Thailand are less likely to become bankruptcy if they are able to effectively generate and manage their profits. To measure the profitability of firms, this study employed a ratio of earnings after taxes to total assets as a proxy, consistent with Beaver (1966).Moreover, the negative relationship between a ratio of earnings after taxes to total assets and a chance of bankruptcy is expected since it is reasonable to argue that with low profits, firm would have difficulty to survive during the bad year of losses due to their inefficient use of operating assets in continuously generating profits. 3. Methodology Data Collection This study emphasizes on Small and Medium Enterprises (SMEs) in form of company limited in Thailand during For sample selection, firstly, the bankrupt firms were selected. These bankrupt firms were those were classified by the Business Online Public Company Limited (BOL) as bankruptcy. Basically, they were the firms that went into the process of liquidation and did not continue their business. Next, using one-to-one matching method suggested by Altman (1968), the bankrupt firm was paired with the non-bankrupt one that having about the same size. Nonetheless, this study divided the data into two groups which were in-the-sample data and out-of-sample data. In-the-sample data was the data during and was used to estimate and develop the prediction model. On the other hand, the out-of-sample data was the data during 2009 and 2010 and was used for the purpose of testing model accuracy.table 3below presents the number of bankrupt and non-bankrupt firms used for model creation and accuracy testing purposes. Methodology Table3: Number of Bankrupt and Non-bankrupt Firms Model Creation ( ) Accuracy Testing ( ) Bankruptcy Non- Non- Total Bankruptcy Bankruptcy Bankruptcy Total In this study, logistic regression analysis is the most appropriate statistical technique because this methodology is suitable for the dependent variable which is a binary variable, e.g. bankruptcy and nonbankruptcy, and the independent variables which are metric variables (Suntraruk, 2010).The following is the logistic regression model. Y i = β 0 + β 1 NWTA i + β 2 TLTA i + β 3 TATO i + β 4 EATA i + ε i Where, Y = a binary variable where 1 = bankrupt firm and 0 = otherwise, β 0 is intercept, β n (n=1,2,3,4) presents the coefficients of independent variables, NWTA is a ratio of net working capital to total assets, TLTAis a ratio of total liabilities to total assets,tato is total asset turnover ratio, EATAis a ratio of earnings after taxes to total assets, and ε is the error term. To develop the model, backward stepwise logistic regression was employed by using data during With this statistical method, the prediction model includes only significant variables. Later, this estimated model was used for the bankruptcy prediction for the year To estimate the probability of bankruptcy, the following function is used (Suntraruk, 2010). 792

6 Probability of bankruptcy Y e e X... X Y X... X Next, the percentage of predictive accuracy was calculated to measure the ability of the estimated model in predicting the chance of bankruptcy. The cut-off point of 0.5 was then used to classify the bankrupt firms from those who were not n n n n 4. Empirical Results Model Estimation Table 4 resents the backward stepwise logistic regression results. Initially, all independent variables; a ratio of net working capital to total assets (NWTA), a ratio of total liabilities to total assets (TLTA), total asset turnover ratio (TATO), a ratio of earnings after taxes to total assets(eata), were included in the regression model. The results from table 4 indicate that for the first step, total asset turnover ratio (TATO) and a ratio of earnings after taxes to total assets (EATA)are statistically significant at the 0.01 (p< 0.01) and 0.1(p< 0.10) confidence level, respectively. On the other hand, it is found that a ratio of net working capital to total assets (NWTA), a ratio of total liabilities to total assets (TLTA), are statistically insignificant at the 0.10 confidence level (p> 0.10). Although the Hosmer & Leme show model chi-square of pointed out that this model fits with the data, there are insignificant variables included. Therefore, for the next step, a ratio of net working capital to total assets (NWTA), least contributed variable, was excluded from the regression model in order to obtain the new model that is better fit with the data. Stepwise Logistic Regression Results Table 4: Stepwise Logistic Regression Results Variables Step 1 Step 2 Step 3 NWTA (1.000) TLTA (1.001) (1.001) TATO EATA Constant Hosmer&Lemeshow Chi-square * (0.653) *** (0.704) 0.323** (1.382) * (0.649) *** (0.694) 0.326** (1.386) * (0.627) *** (0.711) 0.433* (1.542) In-the-sample data consisting of 55 bankrupt firms and 55 non-bankrupt firms were used to estimate the model parameters through the backward logistic regression analysis (likelihood ratio method). The procedure began with all variables in the model and removed them one by one if they did not contribute enough to the regression model. A number on the first line is the estimated coefficient. An italic number is the p-value of chi-square statistic, and a number in the parenthesis is the odd ratio. The Hosmer & Leme show chi-square is 793

7 Y. Pugpaichit & P. Suntrauk used to test the overall fit of a logistic regression model. A finding of non significance shows that the model adequately fits the data. *, **, *** Statistically significance at the 0.01, 0.05, and 0.10 levels, respectively. After excluding a ratio of net working capital to total assets (NWTA), the insignificant variable, the results show that total asset turnover ratio (TATO) and a ratio of earnings after taxes to total assets (EATA) are statistically significant at the 0.01 (p< 0.01) and 0.1 (p< 0.10) confidence level, respectively. Nonetheless, a ratio of total liabilities to total assets (TLTA)is not statistically significant at 0.10 confidence level (p> 0.10). Therefore, to obtain the best prediction model, this insignificant variable was removed. Once a ratio of total liabilities to total assets (TLTA) was removed, the results reveal that both variables, that total asset turnover ratio (TATO) and a ratio of earnings after taxes to total assets (EATA) are persistent significant at the 0.01 (p< 0.01) and 0.1 (p< 0.10) confidence level, respectively. In addition, the Hosmer & Lemeshow chi-square of with insignificant p-value indicates that this model fits with the data well. As a result, this study concludes that the proper bankruptcy prediction model of SMEs in Thailand is as follow: Y = TATO 0.341EATA Where, Y = a binary variable where 1 = bankrupt firm and 0 = otherwise, TATOis total asset turnover ratio,eatais a ratio of earnings after taxes to total assets. According to the bankruptcy prediction model above, it reveals that total asset turnover is negatively related to probability of bankruptcy, meaning that that an increase in total asset turnover reduces the occurrence of bankruptcy. This finding is consistent with Altman (1968), Springate (1978), Fulmer et al. (1984), and Suntraruk (2010). In addition, a ratio of earnings after taxes to total assets is found to be negatively related to the probability of bankruptcy, consistent with Beaver (1966) and Suntraruk (2010). This implies that when firms having high profitability, a chance of bankruptcy declines. Therefore, to reduce the chance of bankruptcy, the firms in SMEs industry should attempt to increase their ability to manage their assets to generate revenue effectively. High revenue is synonym to high profit. Those profitable SMEs firms will not face any financial problems, particularly during business downturn, and finally are less likely to become bankrupt. Predictive Ability To ensure that the estimated bankruptcy model has ability to predict the chance of bankruptcy in the future, the probability of bankruptcy of SMEs is calculated using the following cumulative probability function:- probabilit y of e bankruptcy 1 e ( TATO 0.341EATA) ( TATO 0.341EATA) Using the out-of-sample data during , results on table 5 indicates that a total of 17 firms or 68% of bankruptcy group was predicted correctly as bankrupt firms. In addition, a total of 15 non-bankrupt firms or 60% of this group was predicted correctly. Overall, a total of 64% of all samples was predicted accurately. Result of Classification Accuracy of Bankruptcy Prediction Model A total of 25 bankrupt firms and 25non- bankrupt firms during were as the out-of-sample sample. To classify a firm, the probability of bankruptcy for each firm was calculated from the cumulative probability function: probality = e ( TATO 0.341EATA 1+e ( TATO 0.341EATA ) Once the probability of bankruptcy is obtained, each firm is classified using a cutoff point. If the estimated probability is more than 0.50, the firm is classified as bankruptcy and if less than 0.50, the firm is classified as non-bankruptcy. 794

8 Table 5: Result of Classification Accuracy of Bankruptcy Prediction Model Observed Predicted Bankruptcy Non-bankruptcy Percentage Correct Bankruptcy Non-bankruptcy Overall Percentage Conclusion Because a number of Small and Medium Enterprises (SMEs)went bankrupt as increased continuously, a bankruptcy prediction model is one of important warning tools that would help decrease the chance of becoming bankrupt. Hence, the purpose of this study is to develop the bankruptcy prediction model of Small and Medium Enterprises (SMEs) business in Thailand. Using backward stepwise logistic regression analysis, this study reveals that the proper bankruptcy prediction model of SMEs business includes two financial ratios which are total asset turnover ratio (TATO) and a ratio of earnings after taxes to total assets (EATA). In addition, using the out-of-sample data, such bankruptcy model is able to classify the bankrupt and non-bankrupt firms correctly with 64% of accurate classification. Moreover, this study reveals that total asset turnover ratio (TATO) and a ratio of earnings after taxes to total assets (EATA) are negatively related to the probability of bankruptcy. This implies that an increase in total asset turnover ratio (TATO) and a ratio of earnings after taxes to total assets (EATA) can lessen the chance of facing insolvency. Hence, it suggests that the efficient management of resources is necessary for the Small and Medium Enterprises (SMEs). An efficient use of assets would help SMEs firms generate profits continuously and ultimately reduce the chance to become bankrupt. Results of this research are beneficial to groups of people as follows. Firstly, management and employees of SMEs firms would use the bankruptcy prediction model to assess financial situation of firms. Particularly, management would apply the results from this study when stimulating the strategic changes within a firm in order to avoid financial problems. Secondly, creditors would consider the findings from this study when examining the ability of SMEs firms to repay debts. This would help lessen risk of default. Next, this study would provide addition information for investors when considering making an investment in the SMEs. Furthermore, this study would help shareholders predict event of bankruptcy in order to reduce the amount loss of their money from investment on these enterprises. Last but not least, this study benefits to credit agencies as they would use a bankruptcy prediction model as one of many analysis tools to help establish credit ranking and reliability of SMEs. For further study, besides financial ratios, corporate governance variables could be applied as important predictors of SMEs bankruptcy as Daily and Dalton (1994) found that the good constituent common stock, board of director quantity, and corporate governance structures of the firm can decrease the chance of becoming bankrupt. References Altman, I. E., Financial ratios, discriminant analysis and the prediction of corporate bankruptcy.the Journal of Finance23(4): Beaver, W. H., 1966.Financial ratios as predictors of failure.journal of Accounting Research 4: Brigham, E. F. andm. C. Ehrhardt, 2005.Financial management: Theory and practice. Ohio:South-Western. 11 th edition. Charitou, A.,E.Neo phytouandc.charalambous, 2004.Predicting corporate failure: Empirical evidence for the UK.European Accounting Review13(3):

9 Y. Pugpaichit & P. Suntrauk Daily, C. M. andd. R.Dalton,1994.Bankruptcy and corporate governance: The impact of board composition and structure.academic of Management Journal 37(6): Deakin, E. B., A discriminant analysis of predictors of business failure.journal of Accounting Research 10(1): Fulmer, J. G., J. E.Moon, T. A.Gavin, andm. J. Erwin, A bankruptcy classification model for small firms.journal of Commercial Bank Lending66(11): Grammatikos, T. andg.gloubos, Predicting bankruptcy of industrial firms in Greece.Spoudai 34(3-4): Ingram, D., The advantages of financial ratios.retrieved August 10, 2012, from: Kantong, J Red flags on financial failure: The case of Thai corporations. Doctoral dissertation.national Institute of Development Administration. Kraus, A. andr. H. Litzenberger, A state-preference model of optimal financial leverage.journal of Finance28(4): Merwin, L. C., Financing small corporation in five manufacturing industries.the National Bureau of Economic Research, Nam, J. andt.jinn, Bankruptcy prediction: Evidence from Korean listed companies during the IMF crisis. Journal of International Financial Management and Accounting11(3): Ohlson, J. A., Financial ratios and the probabilistic prediction of bankruptcy. Journal of Accounting Research 18(1): Springate, G. L. V., Predicting the possibility of failure in a Canadian firm.unpublished M.B.A. research project. Simon Fraser University. Suntraruk, P. (2010). The predictors of financial distress: Evidence from nonfinancial firms listed in Thailand.Doctoral dissertation.assumption University. The Office of Small and Medium Enterprises Promotion (OSMEP).(2011a). Definition of SMEs. Retrieved July 20, 2011 from: The Office of Small and Medium Enterprises Promotion (OSMEP). (2011b).White paper on Small and Medium Enterprises of Thailand in 2010 trends of Retrieved October 5, 2011 from: 20SMEs/White%20Paper%20on%20SMEs%202010%20and%20Trends%202011/Executive-Summary- Eng.pdf. 796